78 Mr. W. C. Baker on tlw Hall Effect in 



(b) The current through the secondary circuit was balanced 

 against the current in the shunt circuit by altering the 

 value of the resistance R until there was no movement of 

 the galvanometer- spot on closing the key. This was done 

 for both directions of the primary current. 



It was found that when a balance had been obtained 

 the primary circuit might be left closed for about 

 one minute before the ther mo magnetic effects in the 

 plate were sufficient to cause the spot to move from 

 its zero. Now as in actual work this current was 

 never left closed for more than a few seconds, at 

 most, it is easily seen how absolutely negligible are 

 errors from these sources. 



(c) The direction of the magnetic field was reversed, and 

 balances again obtained for both directions of the primary 

 current. 



These balances for the two directions of the primary 

 current differed at the very lowest fields. In such 

 cases the value of the transverse current was calcu- 

 lated separately for each direction of the primary 

 current and the mean taken. In the case of the lowest 

 field — of 12 lines per sq.cm. — this difference was 6 per 

 cent, for two cases, and 14 per cent, in the third; but 

 the values obtained as above do not differ by more 

 than 3 per cent. For fields above 100 lines per sq.cm. 

 this difference is less than 1 per cent., and it vanishes 

 entirely above 300 lines per sq.cm. 



As the readings were taken on different nights, I 

 think it is most likely due to some mechanical strain 

 in the plate caused by a difference in the temperature 

 of the whole bath — the coefficients of expansion of the 

 gold and the ebonite or the copper distributing bars 

 being different. At all events, an examination of the 

 plotted results shows that the apparent lowering of 

 the Hall coefficient at very low fields is in all pro- 

 bability due wholly to errors of experiment as the 

 quantities observed were so small as to lie but very 

 little within the limit of sensibility of the apparatus. 



(d) The magnetic field was again measured as above. If 

 any small difference from the previous value was found, 

 the mean was taken to be the strength of field during the 

 experiment. If the difference were large, due to the 

 cells running down or other cause, the whole measurement 

 was rejected. 



The couple on the suspended system of the galvanometer 



